Exposing the Robustness and Vulnerability of Hybrid 8T-6T SRAM Memory Architectures to Adversarial Attacks in Deep Neural Networks.
Abhishek MoitraPriyadarshini PandaPublished in: CoRR (2020)
Keyphrases
- neural network
- buffer overflow
- auto associative
- digital image watermarking
- random access memory
- hybrid intelligent
- data corruption
- security risks
- pattern recognition
- semi fragile watermarking
- hybrid models
- watermarking scheme
- countermeasures
- associative memory
- image watermarking scheme
- memory management
- power consumption
- multilayer perceptron
- security vulnerabilities
- artificial neural networks
- geometric attacks
- memory hierarchy
- denial of service
- back propagation
- neural network model
- neural nets
- main memory
- memory requirements
- watermark embedding
- data transmission
- denial of service attacks
- memory space
- fuzzy logic
- multi layer
- digital signal processors
- low power